The operation finds unique data records. It creates a class for each of them and collects the indexes of the data records which match to the key of class. If the flag em -auto is given, then new classes are named by using a running index. Otherwise, new classes are named by using the names of the fields and the compination of the values in unique data records.
Example: A typical use is to find unique data records related to some key and then summarize corresponding descriptive fields inside unique keys.
... # summarize data related to customer number (custNro) NDA> select keyFr -f boston.chas NDA> uniq -d keyFr -cout custCld NDA> clstat -d boston -c custCld -dout custData -avg ...